Project Summary Secondary education has been hailed as a ?social vaccine? against HIV acquisition, particularly for young women who continue to experience high HIV incidence in sub-Saharan Africa. Education may provide HIV information, health literacy, and economic empowerment, reducing risks of transactional sex. Yet until recently evidence on the relationship between education and HIV risk was limited to correlational studies, vulnerable to confounding. In a recent paper in Lancet Global Health, we evaluated a natural experiment in Botswana, generating some of the first causal evidence linking education levels to reduced HIV infection rates. In 1996, Botswana changed the grade structure of secondary school, shifting grade 10 from senior secondary to junior secondary. Because there are many more junior secondary schools, this policy sharply reduced distance to the nearest grade 10 classroom and resulted in a 0.8-year increase in total years of schooling for affected birth cohorts. Exploiting this exogenous variation in education, we found that birth cohorts entering secondary school in or after 1996 had significantly lower risk of HIV infection when contacted in later population-based biomarker surveys. One additional year of schooling resulting from the reform reduced the 10-year cumulative incidence of HIV from 25% to 18%. Our findings were cited by PEPFAR/USAID in allocating over $100M to support secondary school expansion for HIV prevention, as part of the DREAMS initiative and the recently announced Secondary Education Expansion for Development (SEED) initiative. Building on this foundation, this R03 proposal for ?Secondary Analyses of Existing Datasets? will determine the proximate behaviors (Aim 1) and the distal mechanisms for behavior change (Aim 2) that explain the impact of Botswana?s schooling reform on reduced HIV acquisition. To implement the proposed research, we have compiled a rich array of secondary datasets not available to us in our original study, including all four waves of the Botswana AIDS Impact Survey, four Botswana Censuses, and two Botswana Labor Force Surveys, which provide measures of proximate and distal mechanisms. We will analyze the data using the same econometric approach that we implemented in our original study. Our hypothesis is that secondary schooling shifted fertility and labor market preferences, leading to economic independence and lower HIV risk. There is a critical need to understand the mechanisms through which secondary education reduces HIV risk. The proposed research will shed light on whether the lessons from Botswana can be generalized to other settings and what mechanisms should be targeted by future education reforms for maximum HIV impact.